144 



BULLETIN 61, UNITED STATES NATIONAL MUSEUM. 



there is apparently a ojreater tendency toward a smaller formula than 

 21-19-17 than elsewhere in the range. This reduction is a real one, 

 and it takes but little familiarity with the material to bring out the 

 fact that there is a reduction in the number of scale rows all along 

 the western boundary of the form (see map, fig. 59). It is in part 

 owing to this fact that the number of scale rows of elegans (in the old 

 sense) has been so frequently a disputed subject. 



The supralabials are nearly always 8, but in nearly every locality, 

 irrespective of its place in the range, one or two specimens have 7 

 upon one or, more rarely, both sides, but 9 plates are so rare that we 

 have seen but one specimen which had this number and then only on 

 one side. Likewise the number of infralabials, while 10 in the great 

 majority of specimens, is occasional!}^ 9 and much more rarely 11. In 

 fact the average formula for every locality where a considerable num- 

 ber of specimens have been examined is never higher than 8/1 0, and 

 very often about 7.8/9.8, thus indicating that the form tends to have a 

 smaller formula than 8/10. No geographic differences can be observed. 



As in the case of the labials, the material show^s no geographic dif- 

 ferences in the ventral and subcaudal scutes and tail lenijtli. The 



^X\ 



lC) 



a 



a 



8«) 



50 



Fig. 61. — Diagram showing the variation in the number of preoculars in Thamnophis ordi- 



noides elegans. 



sexual variations are as follows: Ventrals, 152-169 (females), 163-182 

 (males), average about 169; subcaudals, 65-79 (females), 80-96 

 (males); tail length, .18-.24 (females), .24-.276 (males). 



Over the greater part of its range elegans has but a single preocular. 

 As Cope (1883, 21-22) first pointed out, however, certain garter- 

 snakes from southeastern Oregon have two scutes in this series. 

 Cope made this trait the basis of a distinct species (hiscutata), but 

 the likeness of these snakes to elegans (vagrans) were recognized by 

 Brown (1901, 23-24) and Van Denburgh (1S97, 212), and they were 

 classed as a subspecies of elegans (and vagrans). "^ As this is the only 

 character by which these snakes differ from typical elegans, and it is 

 far from being constant in this region, it is absurd to attempt to dis- 

 tinguish a distinct subspecies on this basis. Nevertheless the occur- 

 rence of two preoculars in the form is interesting. In the diagram 

 (fig. 61) I have plotted the distribution of the trait, and there is 

 shown to be a strong tendency toward divided preoculars in northern 



» Under the name biscutata, Brown and Van Denburgh also included with these 

 elegans specimens individuals of ordinoides with two preoculars. 



